COVID-19: Countries Wealth, Cases, Deaths and Vaccines

Motivation \ Coronavirus is one of the common viruses that can cause infection in your sinuses, nose or upper throat. Most of them are not dangerous as they present with mild symptoms and can be treated symptomatically, but it killed around 858 people from Middle East respiratory syndrome (MERS) in 2015, this is because of its severe presentation causing respiratory failure. Undiscovered coronaviruses like the Coronavirus are very dangerous because specific treatment for such viruses is not yet available and it rapidly progresses to cause multiorgan failure. This category of the virus causes harmful diseases in mammals as well as in birds. In humans, the virus causes mild respiratory infections, which in rare cases may even cause death. In animals like cows and pigs, it causes diarrhea, while in chickens it causes severe respiratory infections. You may be shocked to know that there are no vaccines currently available for the treatment of this disease.During the Covid-19 outbreak, lots of amazing dashboards were released. However, I was particularly curious about how the wealth of a country affects the main indicators like the number of cases, deaths and vaccines. Hope this post helps people gain some insight on that as well!

Exploratory Data Analaysis on

1. Cases, Deaths, Vaccinations.
2. Reproduction Rate.
3. Stringency Index.
4. Vaccines used.

Questions

1. Do Wealthy countries have less cases?
2. Do Wealthy countries have less deaths?
3. Do Wealthy countries have more vaccines? Did they have them earlier?

Data Source

Special thanks to Our World in Data for making available such a complete, clean, updated and reliable data as they are doing with such an important topic.And the last date in the dataset is Yesterday(2022/5/13)\ Our World in Data (https://ourworldindata.org/ - https://github.com/owid/covid-19-data/tree/master/public/data)

Importing Libraries

import numpy as np
np.set_printoptions(suppress=True)
import pandas as pd
pd.set_option('display.max_columns', None)
import matplotlib.pyplot as plt
import seaborn as sns
import plotly.express as px
import plotly.graph_objs as go
import missingno as msno
from datetime import datetime
from datetime import date, timedelta
from sklearn.preprocessing import MinMaxScaler

Datasets

# Reading the datasets

# Dataset 1: general compiled dataset (key=Co-Date)
df = pd.read_csv('owid-covid-data.csv')
df['date'] = pd.to_datetime(df['date']).dt.date
df = df[df.iso_code.apply(lambda x: x[:5]!='OWID_')].copy()

# Dataset 2: vaccinations dataset for quantity of vaccines (key=Co-Date)
df_vacc_quantity= pd.read_csv('vaccinations.csv')
df_vacc_quantity.dropna(subset=['iso_code'], inplace=True)
df_vacc_quantity = df_vacc_quantity[df_vacc_quantity.iso_code.apply(lambda x: x[:5]!='OWID_')]

# Dataset 3: vaccinations currently used by each country (key=Co)
df_vacc_type = pd.read_csv('locations.csv')
#df_vacc_type = df_vacc_type[df_vacc_type.iso_code.apply(lambda x: x[:5]!='OWID_')].copy()

# Dataset 4: vaccinations by manufacturer (key=Co-Date-Vacc)
df_vacc_manuf = pd.read_csv('vaccinations-by-manufacturer.csv')

key_check = len(df)==len(df[['iso_code', 'continent', 'location', 'date']].drop_duplicates())
print('OVERALL FIGURES')
print('Total countries: \t',len(df.location.unique()))
print('First date: \t\t', min(df.date).strftime('%Y-%m-%d'))
print('Last date: \t\t', max(df.date).strftime('%Y-%m-%d'))
print('Is key Country-Date?: \t', key_check)
print('Shape of the df table is:',df.shape)

OVERALL FIGURES
Total countries: 	 228
First date: 		 2020-01-01
Last date: 		 2022-05-13
Is key Country-Date?: 	 True
Shape of the df table is: (173946, 67)

We know that we have a total of 220 countries.\ The first date in the dataset is January 1st of 2020.\ The last date in the dataset is Yesterday\ The key is Country-Date, meaning that there is only 1 row for each day and country.\ More info about the columns of the dataset can be found in: \ https://github.com/owid/covid-19-data/blob/master/public/data/owid-covid-codebook.csv \ https://github.com/owid/covid-19-data/tree/master/public/data

I will now split the df in two dataframes, one country related and another one country&date. Meaning that one will be used for data particularly related with covid evolution and the other one with parameters from the country itself.

Country indicators

# Defining df_co: dataframe with country related variables (independent of date)
co_related = ['iso_code', 'continent', 'location', 'population', 'population_density',
              'gdp_per_capita','life_expectancy', 'human_development_index']
df_co = df[co_related].drop_duplicates().reset_index(drop=True)
df_co.head(3)

	iso_code	continent	location	population	population_density	gdp_per_capita	life_expectancy	human_development_index
0	AFG	Asia	Afghanistan	39835428.0	54.422	1803.987	64.83	0.511
1	ALB	Europe	Albania	2872934.0	104.871	11803.431	78.57	0.795
2	DZA	Africa	Algeria	44616626.0	17.348	13913.839	76.88	0.748

# Quick look at missing values
msno.matrix(df_co);

From this plot we can see that we have some missing values in the population_density, gdp_per_capita, life_expectancy and human_development_index but no so many. Meaning that this dataframe, the variables per countries are quite complete.

# Check countries with at least 1 missing value
df_co[df_co.isnull().any(axis=1)]

	iso_code	continent	location	population	population_density	gdp_per_capita	life_expectancy	human_development_index
3	AND	Europe	Andorra	77354.0	163.755	NaN	83.73	0.868
5	AIA	North America	Anguilla	15125.0	NaN	NaN	81.88	NaN
9	ABW	North America	Aruba	107195.0	584.800	35973.781	76.29	NaN
21	BMU	North America	Bermuda	62092.0	1308.820	50669.315	82.59	NaN
24	BES	North America	Bonaire Sint Eustatius and Saba	26445.0	NaN	NaN	77.79	NaN
28	VGB	North America	British Virgin Islands	30423.0	207.973	NaN	79.07	NaN
37	CYM	North America	Cayman Islands	66498.0	256.496	49903.029	83.92	NaN
45	COK	Oceania	Cook Islands	17572.0	NaN	NaN	76.25	NaN
49	CUB	North America	Cuba	11317498.0	110.408	NaN	78.80	0.783
50	CUW	North America	Curacao	164796.0	362.644	NaN	78.88	NaN
66	FRO	Europe	Faeroe Islands	49053.0	35.308	NaN	80.67	NaN
67	FLK	South America	Falkland Islands	3528.0	NaN	NaN	81.44	NaN
71	PYF	Oceania	French Polynesia	282534.0	77.324	NaN	77.66	NaN
77	GIB	Europe	Gibraltar	33691.0	3457.100	NaN	79.93	NaN
79	GRL	North America	Greenland	56868.0	0.137	NaN	71.70	NaN
81	GUM	Oceania	Guam	170184.0	304.128	NaN	80.07	NaN
83	GGY	Europe	Guernsey	63385.0	NaN	NaN	NaN	NaN
97	IMN	Europe	Isle of Man	85410.0	147.872	NaN	81.40	NaN
102	JEY	Europe	Jersey	101073.0	NaN	NaN	NaN	NaN
115	LIE	Europe	Liechtenstein	38254.0	237.012	NaN	82.49	0.919
118	MAC	Asia	Macao	658391.0	20546.766	104861.851	84.24	NaN
131	MCO	Europe	Monaco	39520.0	19347.500	NaN	86.75	NaN
134	MSR	North America	Montserrat	4981.0	NaN	NaN	74.16	NaN
139	NRU	Oceania	Nauru	10873.0	682.450	12895.635	59.96	NaN
142	NCL	Oceania	New Caledonia	288217.0	15.342	NaN	77.55	NaN
147	NIU	Oceania	Niue	1614.0	NaN	NaN	73.71	NaN
149	MNP	Oceania	Northern Mariana Islands	57910.0	119.878	NaN	76.74	NaN
160	PCN	Oceania	Pitcairn	47.0	NaN	NaN	NaN	NaN
163	PRI	North America	Puerto Rico	2828246.0	376.232	35044.670	80.10	NaN
168	SHN	Africa	Saint Helena	6095.0	NaN	NaN	80.56	NaN
171	SPM	North America	Saint Pierre and Miquelon	5771.0	NaN	NaN	81.07	NaN
174	SMR	Europe	San Marino	34010.0	556.667	56861.470	84.97	NaN
182	SXM	North America	Sint Maarten (Dutch part)	43421.0	1209.088	36327.232	78.95	NaN
186	SOM	Africa	Somalia	16359500.0	23.500	NaN	57.40	NaN
189	SSD	Africa	South Sudan	11381377.0	NaN	1569.888	57.85	0.433
196	SYR	Asia	Syria	18275704.0	NaN	NaN	72.70	0.567
197	TWN	Asia	Taiwan	23855008.0	NaN	NaN	80.46	NaN
203	TKL	Oceania	Tokelau	1368.0	NaN	NaN	81.86	NaN
209	TCA	North America	Turks and Caicos Islands	39226.0	37.312	NaN	80.22	NaN
210	TUV	Oceania	Tuvalu	11925.0	373.067	3575.104	67.57	NaN
216	VIR	North America	United States Virgin Islands	104218.0	306.480	NaN	80.58	NaN
220	VAT	Europe	Vatican	812.0	NaN	NaN	75.12	NaN
223	WLF	Oceania	Wallis and Futuna	11094.0	NaN	NaN	79.94	NaN
224	ESH	Africa	Western Sahara	611872.0	NaN	NaN	70.26	NaN

Country-Date cases, deaths, tests, vaccinations, reproduction rate, stringency index

# Quick look at missing values in df
msno.matrix(df, labels=True);

Thanks to this plot, we can easily se that there are a lot of missing values for ICU/hospitalized patience. This may be explained due to several countries not reporting it. That is why those columns will not be used further in this notebooks and will be dropped straight away. \ Similarly, we can see high number of missing values in the columns related with the vaccinations. In this case, the reason is not the same as before, instead this could be explained due to vaccines only available in 2021, but covid started since 2019/20. \ Furthermore, we can also see missing values in the tests columns. This can be explained due to two factors, some countries did not report the tests performed at all, and some other reported in a less frequent base than daily (for example, weekly).

# Defining df_co_dt: dataframe with country AND date related variables (date dependant)
# Droppping all ICU because they have lot of missing values and it is not scope of this notebook.
co_dt_related = ['iso_code', 'continent', 'location', 'date', 
                 'total_cases', 'new_cases','total_deaths', 'new_deaths', 'total_tests', 'new_tests',
                 'total_cases_per_million','new_cases_per_million','total_deaths_per_million', 
                 'new_deaths_per_million','total_tests_per_thousand', 'new_tests_per_thousand',
                 'total_vaccinations','people_vaccinated', 'people_fully_vaccinated', 'new_vaccinations',
                 'total_vaccinations_per_hundred','people_vaccinated_per_hundred', 'people_fully_vaccinated_per_hundred',
                 'reproduction_rate', 
                 'stringency_index']
df_co_dt = df[co_dt_related].drop_duplicates().reset_index(drop=True)
df_co_dt.head(3)

	iso_code	continent	location	date	total_cases	new_cases	total_deaths	new_deaths	total_tests	new_tests	total_cases_per_million	new_cases_per_million	total_deaths_per_million	new_deaths_per_million	total_tests_per_thousand	new_tests_per_thousand	total_vaccinations	people_vaccinated	people_fully_vaccinated	new_vaccinations	total_vaccinations_per_hundred	people_vaccinated_per_hundred	people_fully_vaccinated_per_hundred	reproduction_rate	stringency_index
0	AFG	Asia	Afghanistan	2020-02-24	5.0	5.0	NaN	NaN	NaN	NaN	0.126	0.126	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	8.33
1	AFG	Asia	Afghanistan	2020-02-25	5.0	0.0	NaN	NaN	NaN	NaN	0.126	0.000	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	8.33
2	AFG	Asia	Afghanistan	2020-02-26	5.0	0.0	NaN	NaN	NaN	NaN	0.126	0.000	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	8.33

# Quick look at missing values
msno.matrix(df_co_dt);

ww_cases = df_co_dt.new_cases.sum()
ww_deaths = df_co_dt.new_deaths.sum()
ww_vaccinations = df_co_dt.new_vaccinations.sum()
print('Current acumulated Worldwide figures:')
print('Cases:\t\t',f'{int(ww_cases):,}')
print('Deaths:\t\t',f'{int(ww_deaths):,}')
print('Vaccinations:\t',f'{int(ww_vaccinations):,}')

Current acumulated Worldwide figures:
Cases:		 519,287,816
Deaths:		 6,215,905
Vaccinations:	 9,731,621,410

fig, ax = plt.subplots(1, 2, figsize=(15, 4))
sns.barplot(x=[ww_cases, ww_deaths, ww_vaccinations], y=['Cases','Deaths', 'Vaccinations'], 
            ax=ax[0]).set_title('Worldwide acumulated figures')
#check notation of xticks
ax[1].pie(x=[ww_cases,ww_deaths], labels=['Cases','Deaths'], autopct='%.2f%%',
          explode=(0, 0.3), wedgeprops=dict(width=0.7))
plt.title('Worldwide deaths over cases')
plt.show()

This plots are quite simple, but provides a quick overview of the current situation of Covid. Where we can see that vaccines, finally, vaccinations have overcame - by far - the quantity of cases. \ Parallelly, the pie chart on the right hand, show small porcentage (yet relevant and important) of deaths among positive cases reported.

Vaccinations

df_vacc = df_vacc_quantity.merge(df_vacc_type[['location','vaccines']], how='left').merge(
    df_co[['location','continent', 'population', 'population_density',
           'gdp_per_capita', 'life_expectancy', 'human_development_index']], how='left')
df_vacc['date'] = pd.to_datetime(df_vacc['date']).dt.date
df_vacc.info()

<class 'pandas.core.frame.DataFrame'>
Int64Index: 92956 entries, 0 to 92955
Data columns (total 23 columns):
 #   Column                               Non-Null Count  Dtype  
---  ------                               --------------  -----  
 0   location                             92956 non-null  object 
 1   iso_code                             92956 non-null  object 
 2   date                                 92956 non-null  object 
 3   total_vaccinations                   44452 non-null  float64
 4   people_vaccinated                    41984 non-null  float64
 5   people_fully_vaccinated              39654 non-null  float64
 6   total_boosters                       17793 non-null  float64
 7   daily_vaccinations_raw               35416 non-null  float64
 8   daily_vaccinations                   92635 non-null  float64
 9   total_vaccinations_per_hundred       44452 non-null  float64
 10  people_vaccinated_per_hundred        41984 non-null  float64
 11  people_fully_vaccinated_per_hundred  39654 non-null  float64
 12  total_boosters_per_hundred           17793 non-null  float64
 13  daily_vaccinations_per_million       92635 non-null  float64
 14  daily_people_vaccinated              91723 non-null  float64
 15  daily_people_vaccinated_per_hundred  91723 non-null  float64
 16  vaccines                             92956 non-null  object 
 17  continent                            92956 non-null  object 
 18  population                           92956 non-null  float64
 19  population_density                   88411 non-null  float64
 20  gdp_per_capita                       82916 non-null  float64
 21  life_expectancy                      91985 non-null  float64
 22  human_development_index              81367 non-null  float64
dtypes: float64(18), object(5)
memory usage: 17.0+ MB

#Top Countries on Total Vaccinations
n=20
data = df_vacc.groupby('location').max().sort_values(by='total_vaccinations', ascending=False).reset_index()\
            [['location','total_vaccinations','gdp_per_capita']][:n]
fig = px.bar(data, 
             x="location", 
             y="total_vaccinations", 
             orientation='v' 
#             ,color='greens'                
#             ,color_continuous_scale = 'blues'
             ,title='<b>Top {} countries in Total Vaccinations</b>'.format(n)
            )
fig.update_layout(yaxis={'categoryorder':'category descending'})
fig.update_traces(marker_color='green')
fig.update_layout({'plot_bgcolor': 'rgba(0, 0, 0, 0)','paper_bgcolor': 'rgba(0, 0, 0, 0)'})
fig.show()

This plot tell us the countries that have applied the most quantity of vaccines to date. It is biased due to the population of each of them. The bigger the country, the more vaccinations, but not necessarily is the one that is doing that best.

#Top Countries on Total Vaccinations per Hundred
n=20
data = df_vacc.groupby('location').max().sort_values(by='total_vaccinations_per_hundred', ascending=False).reset_index()\
            [['location','total_vaccinations_per_hundred','gdp_per_capita']][:n]
fig = px.bar(data, 
             x="location", 
             y="total_vaccinations_per_hundred", 
             orientation='v' 
#             ,color='gdp_per_capita'                
#             ,color_continuous_scale = 'blues'
             ,title='<b>Top {} countries on Total Vaccinations per Hundred</b>'.format(n)
            )
fig.update_layout(yaxis={'categoryorder':'category descending'})
fig.update_traces(marker_color='green')
fig.update_layout({'plot_bgcolor': 'rgba(0, 0, 0, 0)','paper_bgcolor': 'rgba(0, 0, 0, 0)'})
fig.show()

Similar to previous plot, but this time it is normalized by the population of the country. So one can really see which countries are doing it best (possibly smaller countries find it easier/faster to vaccinate the whole population).

#Map - Total Vaccionations per Hundred
data = df_vacc.groupby(by='location').max().reset_index()[['location','total_vaccinations_per_hundred']]

fig = px.choropleth(data,
                    locations = 'location', 
                    locationmode = 'country names',
                    color = 'total_vaccinations_per_hundred',
                    title = '<b>Total Vaccinations per Hundred</b>',
                    color_continuous_scale = 'greens',
                    hover_name = 'location'
                    ,labels={'total_vaccinations_per_hundred':'% Vacc'}
                   )
fig.show()

Now we can see the same information as in the previous plot, but in a map which makes it easy have all the information in one view.

#Top Countries on Daily Vaccinations per Million in the last 7 days
days_to_subtract = 7
dt_today = date.today()
dt_since = dt_today - timedelta(days=days_to_subtract)
data = df_vacc[df_vacc.date>=dt_since]

#Top Countries on Daily Vaccinations per million
n=20
data = df_vacc.groupby('location').max().sort_values(by='daily_vaccinations_per_million', ascending=False).reset_index()\
            [['location','daily_vaccinations_per_million','gdp_per_capita']][:n]
fig = px.bar(data, 
             x="location", 
             y="daily_vaccinations_per_million", 
             orientation='v' 
#             ,color='gdp_per_capita'                
#             ,color_continuous_scale = 'blues'
             ,title='<b>Top {} countries with more Daily Vaccinations per Million in the last {} days</b>'.format(n, days_to_subtract)
            )
fig.update_layout(yaxis={'categoryorder':'category descending'})
fig.update_traces(marker_color='green')
fig.update_layout({'plot_bgcolor': 'rgba(0, 0, 0, 0)','paper_bgcolor': 'rgba(0, 0, 0, 0)'})
fig.show()

From this plot we can see which countries are currently (from last 7 days) vaccinating at a higher pace, considering the population of the country.

#Map - Daily Vaccionations per Hundred
days_to_subtract = 7
dt_today = date.today()
dt_since = dt_today - timedelta(days=days_to_subtract)
data = df_vacc[df_vacc.date>=dt_since]

data = data[['location','daily_vaccinations_per_million']].groupby(by=['location']).max().reset_index()
#data['daily_vaccinations_per_hundred'] = round((data['daily_vaccinations']*100) / data['population'],4)

fig = px.choropleth(data,
                    locations = 'location', 
                    locationmode = 'country names',
                    color = 'daily_vaccinations_per_million',
                    title = '<b>Daily Vaccinations per Million in the last {} days</b>'.format(days_to_subtract),
                    color_continuous_scale = 'greens',
                    hover_name = 'location'
                    ,labels={'daily_vaccinations_per_million':'Vacc / M'}
                   )
fig.show()

From this map we can see which countries are vaccinating more, considering the lasts 7 days and the population of the countries.

Scatters

def plot_custom_scatter(df, x, y, size, color, hover_name, title):
    '''
    INPUT:
        df - Pandas dataframe - contains the data
        x - string - name of the column to be used as horizontal axis
        y - string - name of the column to be used as vertical axis
        size - string - name of the column to be used to define the size of the bubbles
        color - string - name of the column to be used to define the color of the bubbles
        hover_name - string - name of the column to be used to display when hover on top of the bubbles
        title - string - title of the plot
    OUTPUT:
        plot
        
    Description: the function plots a custom scatter plot with Plotly Express
    '''
    fig = px.scatter(df, x=x, y=y, size=size, color=color,
               hover_name=hover_name, size_max=80, title = title
#                     ,width=800
                     ,height=800
                    )
    fig.update_layout({'legend_orientation':'h'})
    fig.update_layout(legend=dict(yanchor="top", y=-0.1))
    fig.update_layout({'legend_title':'Continent'})
    fig.update_layout({'plot_bgcolor': 'rgba(0, 0, 0, 0)','paper_bgcolor': 'rgba(0, 0, 0, 0)'})
    fig.update_xaxes(showline=True, linewidth=1, linecolor='black', mirror=True)
    fig.update_yaxes(showline=True, linewidth=1, linecolor='black', mirror=True)
    fig.update_xaxes(zeroline=True, zerolinewidth=0.5, zerolinecolor='grey')
    fig.update_yaxes(zeroline=True, zerolinewidth=0.5, zerolinecolor='grey')
    fig.update_xaxes(showgrid=True, gridwidth=1, gridcolor='lightgrey')
    fig.update_yaxes(showgrid=True, gridwidth=1, gridcolor='lightgrey')
    fig.update_traces(marker=dict(symbol='circle-dot',opacity=0.5, line=dict(width=1,color='Black')))
    fig.show()    

# Scatter Vaccination Acumulated Total vs Porcentual, Color per Continent
data = df_vacc.groupby('location').max().reset_index()[['location','continent','total_vaccinations',
       'total_vaccinations_per_hundred', 'vaccines']].dropna()

plot_custom_scatter(data, 
                    x = 'total_vaccinations', 
                    y = 'total_vaccinations_per_hundred', 
                    size = 'total_vaccinations', 
                    color = 'continent',
                    hover_name = 'location', 
                    title = '<b>Vaccinations Acum (Total vs Percent), per continent</b>')

From this plot we can see the relation between the total vaccinations and the total vaccinations considering the population of the country. We can't identify a relation with the continents. It's mainly descriptive, but not much to infer from this plot.

# Scatter Vaccination Rate Total vs Porcentual, Color per type of vaccine
days_to_subtract = 7
dt_today = date.today()
dt_since = dt_today - timedelta(days=days_to_subtract)
data = df_vacc[df_vacc.date>=dt_since]

data = data[['location','continent','total_vaccinations','daily_vaccinations','vaccines','population']].groupby(by='location').max().reset_index().dropna()
data['total_vaccinations_per_million'] = round((data['total_vaccinations']*1000000) / data['population'],0)
data['daily_vaccinations_per_million'] = round((data['daily_vaccinations']*1000000) / data['population'],0)

plot_custom_scatter(data, 
                    x = 'total_vaccinations_per_million', 
                    y = 'daily_vaccinations_per_million', 
                    size = 'total_vaccinations_per_million', 
                    color = 'continent',
                    hover_name = 'location', 
                    title = '<b>Vaccinations Rate (Total vs Daily per million), per continent</b>')

Since the colors represent the continents, we can identify certain aligment withing the European country, meaning that countries in Europe seems to be vaccinating at a relatively similar pace when it comes to daily vaccinations, as well as the accumulated vaccinations. \ We can also identify that African countries seem to be in the bottom left of the chart, which is unfortunately low vaccinations accumulated and daily (except Seychelles). \ Apart from this two observations, the rest of the datapoints seem dispersed to draw conclusions only with it.

data = df_vacc[['continent','location','total_vaccinations','vaccines']].groupby(by=['continent','location','vaccines']).max().reset_index()
data['world'] = 'World'
data

fig = px.treemap(data, 
                 path = ['world', 'vaccines', 'location'], 
                 values = 'total_vaccinations',
#                 hover_name = 'location',
#                 color = 'gdp_per_capita',
                 color_continuous_scale = 'RdBu',
                 title = '<b>Total vaccinations per schema</b>'
                 ,color_discrete_sequence = px.colors.qualitative.Pastel
#                ,labels={'gdp_per_capita':'GDP per capita'}
                )

fig.show()

With this treemap, we can see how much a vaccination schema is used in total. This treemap is a bit misleading due to few countries reporting how many dosis per vaccine were applied (per day or total), hence if a country is applying 10% of a certain vaccine and 90% of another one (supposing only 2 for this example), by looking at the plot you may thing that it is applying 50% of each when that's not the case.

data = df_vacc[['continent','location','total_vaccinations_per_hundred','vaccines']].groupby(by=['continent','location','vaccines']).max().reset_index()
data['world'] = 'World'

fig = px.treemap(data, 
                 path = ['world', 'vaccines', 'location'], 
                 values = 'total_vaccinations_per_hundred',
#                 hover_name = 'location',
#                 color = 'gdp_per_capita',
                 color_continuous_scale = 'RdBu',
                 title = '<b>Total vaccinations per million per schema</b>'
                 ,color_discrete_sequence = px.colors.qualitative.Pastel
#                ,labels={'gdp_per_capita':'GDP per capita'}
                )

fig.show()

This plot is similar to the one before, but considering the population of the countries

Cases

data = df[['iso_code','continent','location','total_cases','population']].groupby(by=['iso_code','continent','location']).max().reset_index()
data['total_cases_per_million'] = round((data['total_cases']*1000000) / data['population'],0)

fig = px.choropleth(data,
                    locations = 'location', 
                    locationmode = 'country names',
                    color = 'total_cases_per_million',
                    title = '<b>Total cases per million habitants</b>',
                    color_continuous_scale = 'reds',
                    hover_name = 'location'
                    ,labels={'total_cases_per_million':'Total cases per million'}
                   )
fig.show()

Here we can have an overall view of the covid cases per million habitant. We can see highly affected areas in South America, Europe and USA.

days_to_subtract = 7
dt_today = date.today()
dt_since = dt_today - timedelta(days=days_to_subtract)
data = df[df.date>=dt_since]

data = data[['iso_code','continent','location','new_cases','population']].groupby(by=['iso_code','continent','location']).max().reset_index()
data['new_cases_per_million'] = round((data['new_cases']*1000000) / data['population'],0)

fig = px.choropleth(data,
                    locations = 'location', 
                    locationmode = 'country names',
                    color = 'new_cases_per_million',
                    title = '<b>Daily new cases per million habitants</b>',
                    color_continuous_scale = 'reds',
                    hover_name = 'location'
                    ,labels={'new_cases_per_million':'Daily cases per million'}
                   )
%matplotlib inline
fig.show()

This plot pretends to show which are the countries where currently present higher daily cases.

Evolution over Time of tests, cases, deaths and vaccinations

data = df.groupby('date').sum()[['new_tests_smoothed','new_cases_smoothed',
                                 'new_deaths_smoothed','new_vaccinations_smoothed']]
data.columns= ['Tests','Cases','Deaths','Vaccinations']
fig = px.line(data, color_discrete_sequence=['grey','blue','red','green'], 
              title='<b>Worldwide daily evolution of Tests, Cases, Deaths and Vaccinations</b>')
fig.show()

With this plot we can see the overall figures evolution across time.

GDP, Treemaps

data = df[['iso_code','continent','location','total_cases','gdp_per_capita','population']].groupby(by=['iso_code','continent','location']).max().reset_index()
data['total_cases_per_million'] = round((data['total_cases']*1000000) / data['population'],0)
data['world'] = 'World'

fig = px.treemap(data, 
                 path = ['world', 'continent', 'location'], 
                 values = 'total_cases_per_million',
#                 hover_name = 'location',
                 color = 'gdp_per_capita',
                 color_continuous_scale = 'RdBu',
                 title = '<b>Total cases per Million and GDP per capita</b>'
                ,labels={'gdp_per_capita':'GDP per capita'}
                )

fig.show()

From this coloured treemap, I was expecting to see some relation with the GDP, accumulated cases and the continents but it is not the case. There is not trend I can highlight.

days_to_subtract = 7
dt_today = date.today()
dt_since = dt_today - timedelta(days=days_to_subtract)
data = df[df.date>=dt_since]

data = data[['iso_code','continent','location','new_cases','gdp_per_capita','population']].groupby(by=['iso_code','continent','location']).max().reset_index()
data['daily_cases_per_million'] = round((data['new_cases']*1000000) / data['population'],0)
data['world'] = 'World'
#data = data[data.new_cases.notnull()]
#data=data.dropna()

fig = px.treemap(data, 
                 path = ['world', 'continent', 'location'], 
                 values = 'daily_cases_per_million',
#                 hover_name = 'location',
#                 color = 'gdp_per_capita',
#                 color_continuous_scale = 'RdBu',
                 title = '<b>Daily cases per million</b>'#and GDP per capita</b>'
                 ,labels={'gdp_per_capita':'GDP per capita'}
                 ,color_discrete_sequence = px.colors.qualitative.Pastel
                )

fig.show()

This treemap, shows which countries are currently having more daily cases, grouped by continents. It is similar to the "Daily new cases per million habitants" map, but in here we can see better the scales. In the map, small countries could hardly be seen in a color scale, while here its size is easier to understsand and call attention.

Stringency Index

This is a composite measure based on nine response indicators including school closures, workplace closures, and travel bans, rescaled to a value from 0 to 100 (100 = strictest). If policies vary at the subnational level, the index is shown as the response level of the strictest sub-region.

data = df_co_dt[['iso_code','continent','location','date','stringency_index']].copy()
data['date'] = pd.to_datetime(data['date']).dt.strftime('%m/%d/%Y')
#data
fig = px.choropleth(data, 
                    locations = 'location',
                    color = 'stringency_index', 
                    animation_frame = 'date',
                    color_continuous_scale = 'blues',
                    locationmode = 'country names',
                    range_color = [0, 100],
                    title = '<b>Stringency Index variation over time</b>',
                    hover_name = 'location'
                    ,labels={'stringency_index':'Stringency'}
                    )
#fig.update_layout(transition = {'duration': 5})
fig.layout.updatemenus[0].buttons[0].args[1]['frame']['duration'] = 30
fig.layout.updatemenus[0].buttons[0].args[1]['transition']['duration'] = 2

fig.show()

Above animated map, shows how the severity of restrictions evolved in each of the countries.

Reproduction Rate

data = df_co_dt[['iso_code','continent','location','date','reproduction_rate']].copy()
data['date'] = pd.to_datetime(data['date']).dt.strftime('%m/%d/%Y')
#data
fig = px.choropleth(data, 
                    locations = 'location',
                    color = 'reproduction_rate', 
                    animation_frame = 'date',
                    color_continuous_scale = 'reds',
                    locationmode = 'country names',
                    range_color = [0, 3],
                    title = '<b>Reproduction Rate variation over time</b>',
                    hover_name = 'location'
                    ,labels={'reproduction_rate':'Rate'}
                    )
#fig.update_layout(transition = {'duration': 5})
fig.layout.updatemenus[0].buttons[0].args[1]['frame']['duration'] = 30
fig.layout.updatemenus[0].buttons[0].args[1]['transition']['duration'] = 2

fig.show()

Above animated map, shows how the reproduction rate evolved in each of the countries.

Correlation

variables = ['new_cases','new_deaths','new_tests','new_vaccinations']

plt.figure(figsize=(8,6))

data = df_co_dt[['location']+variables].groupby(by='location').sum().merge(df_co, left_on='location',right_on='location')
for variable in variables:
    data[variable+'_per_million'] = (data[variable]*1000000) / data.population
    data.drop(columns=variable, inplace=True)
#data = cases[['cases_per_million','population','population_density','gdp_per_capita','life_expectancy','human_development_index']]
mask = np.triu(np.ones_like(data.corr(), dtype=bool))
sns.heatmap(data.corr(), annot=True, fmt='.2f', mask=mask)#,cmap='viridis')
plt.title('Correlation per million with country statistics')
plt.show()

This heatmap plots the Pearson correlation indexes for the different main indicators. Again, I was epecting to see more cases or deaths in less develop / low-income countries, but that is not the case (or not strong enought to my eyes).

Questions

1. Do Wealthy countries have less cases?
2. Do Wealthy countries have less deaths?
3. Do Wealthy countries have more vaccines? Did they have them earlier?

labels = ['low','lower-middle','upper-middle','high']
bin_data = df_co.copy()
bin_data['gdp_bin'] = pd.qcut(df_co.gdp_per_capita
                              ,q=4
                              ,labels=labels
                             )
for label in labels:
    print('Bucket {}:'.format(label))
    print(bin_data[bin_data.gdp_bin==label].location.values)

Bucket low:
['Afghanistan' 'Bangladesh' 'Benin' 'Burkina Faso' 'Burundi' 'Cambodia'
 'Cameroon' 'Central African Republic' 'Chad' 'Comoros' "Cote d'Ivoire"
 'Democratic Republic of Congo' 'Djibouti' 'Eritrea' 'Ethiopia' 'Gambia'
 'Guinea' 'Guinea-Bissau' 'Haiti' 'Kenya' 'Kiribati' 'Kyrgyzstan'
 'Lesotho' 'Liberia' 'Madagascar' 'Malawi' 'Mali' 'Marshall Islands'
 'Mauritania' 'Micronesia (country)' 'Mozambique' 'Nepal' 'Niger'
 'Papua New Guinea' 'Rwanda' 'Sao Tome and Principe' 'Senegal'
 'Sierra Leone' 'Solomon Islands' 'South Sudan' 'Tajikistan' 'Tanzania'
 'Togo' 'Tuvalu' 'Uganda' 'Vanuatu' 'Yemen' 'Zambia' 'Zimbabwe']
Bucket lower-middle:
['Albania' 'Angola' 'Armenia' 'Belize' 'Bhutan' 'Bolivia'
 'Bosnia and Herzegovina' 'Cape Verde' 'Congo' 'Dominica' 'Ecuador'
 'Egypt' 'El Salvador' 'Eswatini' 'Fiji' 'Georgia' 'Ghana' 'Guatemala'
 'Guyana' 'Honduras' 'India' 'Indonesia' 'Jamaica' 'Jordan' 'Laos'
 'Moldova' 'Mongolia' 'Morocco' 'Myanmar' 'Namibia' 'Nicaragua' 'Nigeria'
 'Pakistan' 'Palestine' 'Paraguay' 'Peru' 'Philippines'
 'Saint Vincent and the Grenadines' 'Samoa' 'South Africa' 'Sri Lanka'
 'Sudan' 'Timor' 'Tonga' 'Tunisia' 'Ukraine' 'Uzbekistan' 'Vietnam']
Bucket upper-middle:
['Algeria' 'Antigua and Barbuda' 'Argentina' 'Azerbaijan' 'Barbados'
 'Belarus' 'Botswana' 'Brazil' 'Bulgaria' 'Chile' 'China' 'Colombia'
 'Costa Rica' 'Croatia' 'Dominican Republic' 'Equatorial Guinea' 'Gabon'
 'Greece' 'Grenada' 'Hungary' 'Iran' 'Iraq' 'Kazakhstan' 'Latvia'
 'Lebanon' 'Libya' 'Malaysia' 'Maldives' 'Mauritius' 'Mexico' 'Montenegro'
 'Nauru' 'North Macedonia' 'Palau' 'Panama' 'Poland' 'Romania' 'Russia'
 'Saint Kitts and Nevis' 'Saint Lucia' 'Serbia' 'Seychelles' 'Suriname'
 'Thailand' 'Turkey' 'Turkmenistan' 'Uruguay' 'Venezuela']
Bucket high:
['Aruba' 'Australia' 'Austria' 'Bahamas' 'Bahrain' 'Belgium' 'Bermuda'
 'Brunei' 'Canada' 'Cayman Islands' 'Cyprus' 'Czechia' 'Denmark' 'Estonia'
 'Finland' 'France' 'Germany' 'Hong Kong' 'Iceland' 'Ireland' 'Israel'
 'Italy' 'Japan' 'Kuwait' 'Lithuania' 'Luxembourg' 'Macao' 'Malta'
 'Netherlands' 'New Zealand' 'Norway' 'Oman' 'Portugal' 'Puerto Rico'
 'Qatar' 'San Marino' 'Saudi Arabia' 'Singapore'
 'Sint Maarten (Dutch part)' 'Slovakia' 'Slovenia' 'South Korea' 'Spain'
 'Sweden' 'Switzerland' 'Trinidad and Tobago' 'United Arab Emirates'
 'United Kingdom' 'United States']

data = bin_data[['location','gdp_bin']].dropna()

fig = px.choropleth(data
                    ,locations = 'location'
                    ,locationmode = 'country names'
                    ,color = 'gdp_bin'
                    ,title = '<b>Countries Wealth</b>'
                    ,color_discrete_sequence=['red','orange','lightblue','darkblue'] 
                    ,hover_name = 'location'
                   )
fig.update_layout({'legend_orientation':'h'})
fig.show()

def wrap_plot_gdp(variable, date_start='2020-01-01'):    
    '''
    INPUT:
        variable - string - name of the column to be used as values for the lineplot (y axis)
        date_start - string - optional - if none, will plot from first data point. 
            If a date, will render only after that date
    OUTPUT:
        plot
        
    Description: the function wrap and plot the data from the df and renders
    a custom line plot with Plotly Express that shows the evolution of a certain
    variable from the dataframe over time. 
    Optionally, a custom date start can be passed to filter only after that date.
    '''
    labels = ['low','lower-middle','upper-middle','high']
    bin_data = df_co.copy()
    bin_data['gdp_bin'] = pd.qcut(df_co.gdp_per_capita
                                  ,q=4
                                  ,labels=labels
                                 )
    data = df[['location','date']+[variable]].merge(bin_data[['location','gdp_bin']])
    data['date'] = pd.to_datetime(data['date'])
    data = data[data.date>date_start]
    data = data.groupby(['date','gdp_bin']).mean().reset_index()
    data = data.pivot(index='date', columns='gdp_bin', values=variable)

    #plot
    fig = px.line(data
                  ,color_discrete_sequence=['red','orange','blue','darkblue'] 
                  ,title='<b>{} vs gdp_bin</b>'.format(variable)
                  ,labels={'value':variable}
                 )
    fig.update_layout({'legend_orientation':'h'})
    fig.show()

1. Do Wealthy countries have less cases?

wrap_plot_gdp(variable='new_cases_smoothed_per_million')

wrap_plot_gdp(variable='total_cases_per_million')

From the above two plots, we can see that the wealthier the country, the earlier it reported Covid19 cases. We can also see that around May2021, there is a drastic reduction on the high-income countries, possible due to the vaccines. From the second chart, we can see that the high-income countries show more quantity of population tested positive.

2. Do Wealthy countries have less deaths?

wrap_plot_gdp(variable='new_deaths_smoothed_per_million')

wrap_plot_gdp(variable='total_deaths_per_million')

When it comes to deaths, we can see a big and early spike for high-income countries with a drastically reduction around July2020. Again, in January2021 they presented more deaths than other lower income countries. Something that caught my attention in the two plots, is the significantly lower deaths in the lowest income countries. A possible explanation could be that the figures do not reflect what actually happened / is happening in reality.

3. Do Wealthy countries have more vaccines? Did they have them earlier?

wrap_plot_gdp(variable='new_vaccinations_smoothed_per_million', date_start='2020-12-01')

wrap_plot_gdp(variable='total_vaccinations_per_hundred', date_start='2020-12-01')

For vaccinations, we can see again higher numbers for high-income countries. It is not only clear that they are having more vaccines than the lower groups, but also that they started to have them earlier. This may be attributable, among other causes, to the fact that some high-income countries were the firsts ones to develop the vaccine (prioritizing their own population), or also that these countries were able to pay higher prices in order to get the shots sooner.

Conclusion

To sum up, wealthier countries reported a higher number of cases, deaths and vaccines. Furthermore, the wealthier the country, the earlier the vaccination processes started.

Disclaimer

Even though the Our World in Data team made a fantastic job gathering and wrangling the data, the conclusions could not reflect the reality due to countries measuring/reporting in different ways, applying different tests criteria’s, etc.